4 ms·

This was fun, I did a live demo at work a few months ago on this, got stuck on getting the model to predict with enough accuracy to be useful then came across t

This was fun, I did a live demo at work a few months ago on this, got stuck on getting the model to predict with enough accuracy to be useful then came across this post, my activation function was slightly off. Most people understand Fizz Buzz, building up and solving the problem with a neural network shows that a bit of ML code can solve the problem without being programmed explicitly to do so, quite a good example.

skrebbel 9y ago
> ML code can solve the problem without being programmed explicitly to do so Umm, no. He implemented all of Fizz Buzz except the loop in his "fizz_buzz_encode" function. Totally explicit.
- ovi256 9y ago
  That's just code to generate the examples, an example of synthetic data. He could just have written down the expected output instead of doing that.
- nl 9y ago
  That’s the joke! He implements it explicitly but only to generate training data.
amelius 9y ago
> Most people understand Fizz Buzz, building up and solving the problem with a neural network shows that a bit of ML code can solve the problem without being programmed explicitly to do so, quite a good example. If you train your model with completely different data than you use for inference, there is absolutely no guarantee that it will work as expected. In this case, they trained with numbers >= 1024, where numbers <= 100 were used for inference.